In a previous study, investigators of IOBA (Valladolid, Spain) has identified three predictive models based on genetic markers. Machine learning methods, as statistical tools, have been used to predict the probability of developing PVR after primary RRD by using subsets of 197 SNPs in 30 candidate genes. Before applying predictive models in clinical practice, it is necessary to determine if they provide realistic estimations. The external validation has two components: reproducibility and transportability (Justice 1999). While reproducibility is a characteristic related to internal validation, transportability (external validation) is related the temporal and geographical validation (Altman 2000). These characteristics are essential for a formula to be applicable. In this project, the transportability (external validation) of the predictive models obtained from our previous study will be analysed among eight centres including the Rotterdam Eye Hospital, and will compare them to select the best one.